Welcome to episode #183 of the Super Data Science Podcast. Here we go!
Let’s lie low with the technical stuff for now.
Here’s an interesting proposition for today: how about we delve into the world of analytics consulting business and analytics in general?
If this sounds right for you, then you’lllove this episode of Super Data Science Podcast because chatting with me today is Dominic Ligot, or ‘Doc’ for short.
About Dominic Ligot
Dominic “Doc” Ligot is the Founder and Chief Technology Officer of CirroLytix Research Services which helps small and medium enterprises achieve better outcomes with the use of applied analytics solutions. Doc has gained nearly 18 years of experience in data-driven decision making, data management, and analytics.
He’s also the Co-Founder and the Chief Risk Officer at UpLoan.ph, a fintech lender, and a Founding Board Member and Head of R&D for the Analytics Association of the Philippines where he promotes data ethics and analytics freelancing.
Tune in as Doc and I talk about how his consulting company CirroLytix started, how’s the current state of data science and analytics in the Philippines, what the 3 verticals of his consulting company are, how important training is for business decision makers, where the field of data science is going, and many more!
So, to start, Doc talked about his company, CirroLytix. It is a small company made up of barely 10 consultants in Manila. As an analytics consulting firm in a third world country, Doc can attest to the significant growth of data science, data analytics, and other emerging technologies. Doc says that industries are gradually realizing the need for data science and data analytics. He also saw data analytics as a helping hand to the country’s workforce since Philippines has been a major destination for outsourcing for many years now.
According to Doc, here’s something you can apply if you’re looking into starting your own analytics business… the 3 verticals or key areas that CirroLytix has been focusing on:
Data Engineering: They help enterprises how to effectively ingest data and digitized it into its most usable form. They guide the clients to use analytics and the right tools to improve their daily operations.
Consulting: Stepping away from the technology, they talk to business executives about what cases are most appropriate for their company. Increase the business value through data-driven tools.
Training: Whether it be in-house or public training, Doc sees them as rich sources of leads. They are able to test products and services. They find target clients and also future talents.
Doc emphasizes that business decision-makers should be equipped with proper knowledge of analytics to arrive at better data-driven business solutions. Decision-makers should be guided on how to use data in a proper way. Business people can’t speak data language in order to effectively transform objectives to actionable plans and strategies. Correct information is vital to pitch better to their high-level executives. Besides, analytics is the indicator… the scorecard for your business.
During the last few minutes, Doc also shares with us where he sees the field of data science is going. Here are the key points he discussed:
1 – Democratization of Knowledge and Skills
2 – Proliferation of Data Science
3 – Machines Working with Humans
Make sure to listen if you’re intrigued with what he thinks about them!
In this episode you will learn:
- How did Cirrolytix start? (06:25)
- Philippines, one of the emerging countries in Asia-Pacific Region, has significantly evolved for the past 10 years. (09:00)
- It has become a major destination for outsourcing. (11:40)
- Local industries in the Philippines are now realizing the need for analytics. (13:00)
- The growth of data and analytics is moving to the same evolutionary cycle of the freelancing industry. (14:00)
- When’s the good time to study business analytics? (17:35)
- You have to have a dedicated focus as business on analytics to achieve growth. (18:45)
- The concept of knowledge compression. (21:30)
- The 3 verticals of CirroLytix Research Services: (24:50)
- Data Engineering (24:50)
- Consulting (26:36)
- Training (27:13)
- “The state of the analytics industry is still quite silent.” (34:25)
- CirroLytix will have Business Analytics Masterclasses that aims to help business decision-makers how to utilize data. (37:58)
- The need for data ethics. (39:15)
- Being knowledgeable about analytics could help decision-makers pitch effectively to their high-level executives. (45:36)
- Doc’s thoughts on where the field of data science is going. (49:52)
Items mentioned in this podcast:
- Business Analytics Masterclass Series | September – October 2018 | Manila, Philippines
- Competing on Analytics: The New Science of Winning by Thomas H. Davenport
- Life 3.0: Being Human in the Age of Artificial Intelligence by Max Tegmark
Kirill Eremenko: This is Episode number 183 with founder and Chief Technology Officer at Cirrolytix, Dominic Ligot. Welcome to the Super Data Science Podcast. My name is Kirill Eremenko, Data Science Coach and lifestyle entrepreneur. Each week we bring you inspiring people and ideas to help you build your successful career in data science. Thanks for being here today, and now let's make the complex simple.
Welcome back to the Super Data Science Podcast, ladies and gentlemen. Today we've got a very interesting episode. We've got Dominic, or for short, Doc Ligot, joining us on the show, and we are talking all about creating businesses in the space of analytics consulting. Dominic is the founder of Cirrolytix, a data science consulting firm in the Philippines, and they are servicing clients and helping them introduce data science. They're conducting trainings in the space of data science, they're conducting consulting projects, and so on, so a very exciting space to be in.
In this podcast you will learn how Dominic got started out. You'll also learn about the space, the environment, the analytics environment in Philippines, but don't fret if you are not in the Philippines yourself, because we actually discuss in the episode how all of this, everything we talk about, is actually applicable to any data science environment, whether it's a city, or a country, and how to see the telltale signs for that.
Interestingly enough about this episode is that normally on the podcast, we try to cover a variety of topics. We try to go in the technical side of things, we try to talk about business, we talk about careers, however in this specific podcast, we don't talk about anything technical, so if you are after technical topics, then this podcast is probably not for you.
This podcast is for you though, if you are considering creating a start up in the space of analytics, or if you might be considering sometime down the track doing so, or getting into the space of analytics consulting, because we got so carried away with the topic, it was such an interesting conversation, we just thought it would be better not to dive into the technical components of the work that Dominic does, and rather specifically focus on the challenges of starting a analytics consulting business, and where the world is going in the space of analytics in general, and the demand for analytics from the industries and businesses.
So a very interesting chat, I personally learned a lot. I can't wait for you to hear it all. So without further ado, I bring to you Dominic Ligot, founder of Cirrolytix. Welcome to the Super Data Science Podcast, ladies and gentlemen, and today we've got a very exciting guest on the show, Dominic Ligot. Dominic, welcome, how are you going?
Dominic Ligot: Hi Kirill, good apart from the not-so-good weather in Manila, but we're all doing fine, we're all nice and dry. Everyone's wet outside, but yeah, excited to be on the podcast.
Kirill Eremenko: It's so great to have you. We were just chatting before the podcast about the Philippines, and how the Philippines is in the peak, or just about to enter the peak of the typhoon season right now. How does that usually go down?
Dominic Ligot: Well yeah, so the usual is floods, trying to avoid water, trying to get from point A to point B. Actually, it's interesting, because I remember talking in another forum about the Philippines having what you call a typhoon economy. So there's a part of the economy that's reliant on typhoons hitting, so that all the reconstruction, and the plumbers, and the carpenters get something done.
Kirill Eremenko: Oh, wow.
Dominic Ligot: It's kind of a weird thing, because there was one year where we had an interestingly low number of typhoons from the average, and that actually hit the GDP a little bit, so there might be some credence to that theory. It's bizarre.
Kirill Eremenko: Wow, that is so, so counterintuitive. Wow. Interesting. Okay, good to know. Yeah, we have one person working in the Philippines at Super Data Science, and whenever you guys get into typhoon season, there's always problems with the internet, and it's always so hard to get in touch. In fact, I know that sometimes people have two internet providers, just as like a backup at home, in case one goes down.
Dominic Ligot: Yup.
Kirill Eremenko: All right.
Dominic Ligot: Absolutely.
Kirill Eremenko: Dominic, so probably first and most important question, very interestingly, as you mentioned, people call you Doc, so I'm probably going to be calling you Doc throughout the podcast, and to prepare our listeners for that, could you tell us the story behind why people call you Doc?
Dominic Ligot: Yeah, no worries. Actually, it's kind of like a normal kind of preamble, like I say, "Hey, I'm Doc. I'm not a doctor." It's always good for a few seconds of laughs. It's actually a school yard thing, so as early as ... I don't know, maybe six years old, people were calling me Doc for no apparent reason. The name stuck. For a brief moment in time, I was actually considering becoming a medical doctor, and when I realized how much blood that was going to be involved, and cadavers, it just wasn't my thing.
Then much later, I think now especially for the data scientists, you do meet a few doctors in terms of PhDs, and that's always interesting. So people keep asking, "So what did you do your PhD in?" And I say, "Well, I'm not really a doctor." It's always a point for conversation.
Kirill Eremenko: Yeah, wow, that's definitely a great icebreaker. "Hello, I'm Doc, but I'm not a doctor." Raises a few questions. All right, well thank you. Let's dive straight into it. For the purposes for our listeners to get to know you a bit better, you're the founder and Chief Technology Officer of Cirrolytix. Can you tell us a bit about the company and what Cirrolytix actually does?
Dominic Ligot: Yeah okay, so Cirrolytix, just so you demystify the name, cirrus clouds, we're all about doing analytics on the cloud, and of course analytics, so Cirrolytix. We're a small company, barely 10 consultants, give or take a couple of freelancers.
I started the company in 2016, so we've been around for going on two years. The inspiration for Cirrolytix actually came about when ... In a past life I actually worked for an IT company, and you know how it is with these big IT vendors, you do meet clients, especially in data and analytics. They need what you're selling, but some of these solutions, especially when you get into the hardcore data warehousing and software can get pretty expensive.
That was actually a heartbreaker for me, especially working in a country like the Philippines, which is still an emerging economy. There are many small and medium enterprises who really need the benefits of data, but they can't afford it. So I said, "Okay, why don't I just do it myself, after going on 20 years, actually in the industry? I might know enough to do my own thing."
And yeah, so far so good. We've been at it for going on two years. Our main clientele are usually medium sized companies, so normally those with less than 100 employees. They span the gamut from retail, e-commerce, product companies, also other consultants, and usually their needs don't stray too far from the norm. They're starting to accumulate data themselves. Not at the level that enterprise companies and big ones ... So data scientists normally don't stray too far from the gigabytes, occasionally a terabyte, but now they're struggling, because it's stuff that doesn't fit on Excel sheets, and now they realize that that data can be useful for running their business.
When they go out and start talking to the IT vendors and the consultants, they get shocked at just how expensive it gets. That's normally what gets us in the door and say, "Hey, you need your data sorted out. You need to start getting your feet wet with machine learning on a simple level, like for your e-commerce company." Those are the companies we go for.
Kirill Eremenko: Very interesting. Like here, I probably want to mention something that's ... also we chatted before the podcast, that so Philippines, very interesting geographical location, very interesting country, especially for people who haven't been to the Philippines, I think we need to paint a bit of a picture of how this country is set up in terms of analytics, why this need is growing. I'll just mention my side of the story, and maybe then you can add in yours.
So I've never been to Manila. I'm really looking forward to going to Manila one day, I heard so many great things. I have been to an island called Cebu and an island called Malapascua, and my experience was that it's very far from civilization, very non-commercial, non-industrial, like I went there for scuba diving and for the nature, the jungle, and those things.
It's kind of like that was my impression of Philippines. But now you're talking about analytics and all this need, and how the data is growing. Tell us a bit about Manila. What kind of city is it, and what kind of ... like these companies that they operate, are the industries developed? Are the companies themselves growing and developed, and is it like a big market in general, a big economy?
Dominic Ligot: Yeah, yeah. So I guess just for context, you are correct. In anywhere else in the Philippines, it's a tropical country, beaches, and jungles, that's pretty much the scene. Then you have a couple of areas. You mentioned one, which is Cebu, and then of course there's the capital, which is Manila, and there's another one further down south called Davao, these three ... I would call them, are really full-fledged metropolises, truly sprawling. Manila in itself at any given time of the day would have anywhere from 10 to 30 million people, so it's really, really big.
I think the big thing about how the Philippines is evolved, especially in the last 10 years, is that it's become a major destination for outsourcing, so call centers, BPOs, KPOs, have been coming here. A big part is because, well, number one, the government situation, the political situation has stabilized somewhat.
So the Philippines of today is a very stable business environment. There's a very strong American influence, everyone speaks English. I think that's been the fundamentals that's brought a lot of outsourcing to the country, so you've got everyone from the big banks, like Bank of America, JP Morgan, to the big IT firms like Accenture, IBM, and Teradata.
They've all set up initially customer service centers here, and that's branched out in the past 10 years to include other things, like technical support, legal and medical transcription, so it's really given the economy a second wind. A lot of it is really dependent now on outsourcing now than anything.
I think it's reaching a new level of maturity, because you basically created a new workforce that's technology savvy, communication expertise is pretty good, that now even the local industries are starting to pick up in terms of, "Hey look, we can use the people that are coming out of these outsourcing centers." Analytics is one of those thing that it's giving the workforce additional opportunities in addition to being an outsourcing hub.
Kirill Eremenko: Gotcha, gotcha. And so you're in a very interesting and lucrative, I would say, position, as long as you know how to take advantage of it, which it looks like you do, that you are in an emerging market, or like ... It is a big city, but in terms of the need for analytics, it is only now realizing the demand, or like the value of analytics, and you as a consultant, you're positioning yourself that you can provide that service, that value. You can add it to the businesses.
I think a lot of our listeners on the podcast, like in different locations, might find themselves in a similar situation. It might not be like it's a country, like it's a different country, like in the Asia-Pacific, or it's some remote location with jungles on one hand, and big cities on the other. It might be somewhere in Europe, or it might be somewhere in the U.S., but if you take those ... if you strip away those ... like the geographical side of things, and you look at the context, it might be exactly the same that your city, this is for the listeners, that your city or maybe even your country as a whole is now only getting to the stage where the industries and the economy in general are seeing value of analytics, and positioning yourself as a person or a company that can provide the value is a great step in growing a business.
So Doc, can you give us a ... you already mentioned how you came up with the idea, you were 20 years in the industry, but what did it take to actually get started? Because ultimately I would see it as quite a challenging thing to start a business and position yourself out there saying that, "Hey, I can provide this service," and getting your first client, and all those things. If you don't mind sharing a bit of that.
Dominic Ligot: Yeah. A lot of it is really just being fortunate enough to be in the proverbial right place at the right time, and when you say right time, alongside the development of let's say offshoring and outsourcing in the country, the state of telecommunications has improved, and that's actually empowered a lot of ... I don't know if you guys have heard of the term, "digital nomads," so we've got people moving in and out. They can do most of their work from home, and the emergence of let's say cloud services has made doing a lot of work that previously involved a lot of technology locally, now you can do it all in the cloud.
It's easier to collaborate now, it's easier to share data, share files, and just the proliferation of a lot of the ... I think information, especially in analytics on the internet, these are kind of the ... all the factors that got in. If I'm going to point to the single most, I think, factor that got me really started, it's that you find a need, and you see it every day.
Like I started when I was still working in IT, companies are now struggling with information, with data, and on the other hand, you've got a little bit of knowledge that you think you can solve that problem. I think in the first instance it starts there. You know, you start businesses not thinking of money, not thinking of capital, and it's best if you start it with a need or a problem to solve.
Of course once you find that need, the other half is can you actually sell it, or can you actually convince people that it's worth paying for? I think that's where a lot of the people who are thinking of getting into businesses, especially analytics businesses, are going to struggle a bit. Even though data and analytics has been around for, I don't know, 20 or 30 years, it's always been a back office thing, so it's always been kind of like in the background. Now it's becoming more of a foreground investment for companies, but there's still a lot of confusion as to, "Okay, what's a good amount to charge? What's a good amount to pay for this stuff?"
That's classical evolution. I mean, web development and the internet started out the same way just in the '90s, no one knew that the internet would be important, and you had all these occupations related to the internet, like web developers, web designers, even graphic designers. They didn't know how to place themselves back in the day. Then now you've got a very rich freelancing industry related to the internet, and every company kind of takes it as a given that you have to be on the web.
So I kind of see data and analytics moving into a similar evolutionary cycle, but it is early days. I would say with a few exceptions in the world, most economies, most countries are still kind of getting into data and analytics as a more formal field.
Kirill Eremenko: So you would say even despite the challenges of convincing clients to buy, you would say that it is a good time to consider starting an analytics business?
Dominic Ligot: Yeah, absolutely. I think the biggest shift, one of many anyway, is that analytics is suddenly not just an IT problem, because back in the day, I'm sure many can relate, when you buy a BI tool, or run a few even Microsoft Excel for the first part, that used to be stuff that the IT department was worried about, just installing it on your PC and getting it out there.
Nowadays, because these tools are very important to business, it's becoming more of a business investment, and that's shifted the conversation a lot. Now you have marketing people, HR people, finance people, concerned about what kind of tools, what kind of analysis they need to put into play. It's no longer possible to do it by hand, it's no longer possible to do it manually.
A lot of the conversation has shifted from purely IT to now business, and I think that's where analytics best thrives, and kind of like the business domain rather than the pure technology discussion.
Kirill Eremenko: Yeah no, that is definitely a good example. I was just thinking of like, there are just some things that you ... it's better to give to the experts, right? Like yes, you can do it in the back end, but then you need to make sure you have a dedicated focus as a business. If you're going to do analytics as part of your back end operations, you got to make sure you have a dedicated focus to analytics and that you're building out the team, you know what you're doing, and you're following all these industry trends and standards.
And innovation as well. Some things might not be standard, some things might be cutting-edge, leading-edge technology, and at the same time, like ... or you could go find a company such as yours, and say, "Okay, how about you guys do it, and I don't have to worry about it," especially even if a big organization is considering to implement analytics as a back end operation, then at the start, it's going to be hard, right? While you're doing that, you don't want to fall behind your competition, and you still want to be on top.
Plus, I'm sure when you guys go into a business, you coach them, you provide insight. You don't just give the analytics, but you also provide insights on how it's done, and what your approach was, what the methodology was. At the end of the day, my thing would be if you can come in and provide a service, that's great, but if you can coach them to do it on their own, I don't think that's a ... that's actually a good thing, right?
For me, I'd feel great if I'm a consultant, I go into a business, and I guess I'm going to lose them as a client, because they'll build out their own internal capability, but I'll feel good that I can actually help a business grow and do that. What are your thoughts on that?
Dominic Ligot: Well yeah, you brought up a very good point. Even back when I was working in an IT company, that's always kind of the conundrum, right? The moment you introduce a solution, the moment you teach a client how to do something, the initial motivation will always be, "Okay, I don't want to be paying a consultant forever. Let's do it ourselves," or, "Let's pay them long enough so that we learn it."
I think that's fair enough. I think it's important to recognize that even analytics itself or if we use the more, I think popular term, which is data science, right? There are levels to look at. I think there's a basic level where everyone needs to generate reports, everyone needs to be able to manage and cleanse data, and it's descriptive analysis, if you want to talk about it across the spectrum.
But then at the same time, given the developments in say not just in technology, but kind of in the types of data, in the types of use cases that have come to fore in the last 10 years, there's also a need to do a little bit of what I would call ... I don't know if this is the proper term, knowledge compression.
So for example, let's take something esoteric like machine learning. Once upon a time, no one cared about it, or only like proper computer scientists and researchers would even think about doing machine learning, and this is like classical machine learning where you're ... not even the deep learning stuff, where you do neural networks and logistic linear regressions, kind of that level of machine learning.
Most businesses wouldn't care about that, but now that you've got such a rich tapestry of data to choose from, the use cases become even more interesting, and the cost of technology has fallen down. Suddenly, PhDs have a job in what would otherwise be a marketing department. That's only been a recent phenomenon, and you don't pick those guys up from the street, you need experts to come in. Even if you did find these talented individuals, retaining them would be costly, and there isn't enough supply of that expertise. I think that's the niche where a lot of analytic consultants such as myself and some start ups can hop in, because you don't need this high-level, PhD level type of machine learning every day, not like we would need to pick a report, right?
But from time to time, you do need these services to gain an edge on the competition. To give you an example, an esoteric machine learning use case object detection, right? So you want to tell if a picture is a cat or a dog. That used to be just the stuff of science experiments, but now with the advents of open source libraries and machine learning, it's now being democratized, you can actually without spending a dime, build your own object detection and image recognition system in your laptop.
But if you don't have ... let's say a guy has been doing computer vision research, or proper big data technology knowledge, and it's tossed out to any individual, that could spell disaster for a business. On the other hand, if you had the right expertise, the right tools, you can spend that in many, many different ways. You can use object detection for security, for instance, for fraud detection, or you can use that to detect inventories on your shelves without having to resort to manual counting.
These are some of the emerging use cases that suddenly people who used to do this stuff just purely for research is now coming into the commercial domain. I think that's a space where at least for the time being, there is a niche for specialized consulting to come in. But again, we don't just do that, we kind of do everything end to end, a full spectrum, so it's just an interesting development that wouldn't have been possible years ago, given that the ... would be hard to come by, and the technology was too expensive, and the data wouldn't be there. But now you've got a lot of these things happening now.
Kirill Eremenko: Okay, gotcha. When you say full spectrum, can you tell us a bit more? What does that mean?
Dominic Ligot: Yeah okay, so Cirrolytix, our basic let's say verticals, would kind of fall into three areas. One area is in the data engineering side, so this is like the boring stuff most companies don't think they need, but they do. So things that range from how to ingest data from your data sources, or from outside, or just getting data digitized into a proper form, storing that. So not quite full scale data warehousing, like the likes of what IBM and Oracle offer, but small scale data warehousing, like the stuff we can do on Amazon, or on Azure, or on a Snowflake instance.
Then moving on up the value chain to business intelligence, machine learning, and analytics. Then on the far end, getting the outputs of these analysis. It is interesting, because without revealing too much, I think this is a gap right now in the data science industry. You've got a lot of people who can do a lot of fancy analysis, a lot of fancy models, fancy charts, et cetera, but in terms of making it friendly for business user, that's kind of still lacking. I think that's where I would say more traditional software development, application development comes in.
So yeah, never mind that you've got a very good, say neural network that can identify potential customers with 98% accuracy, but if you have to run a whole slew of code to do it, your average marketer won't do it, but if you can get them an app that could do it automatically, then that's kind of bridging the last mile. That's kind of like one vertical for us, getting everything from sourcing the data, all the way to trying to get into an app. That's the data engineering vertical.
The second vertical would be more around consulting, so determining what use cases are appropriate for your company. This is less of a technology discussion, more about transformation, more about what kind of use cases do you do? What do I need to do to improve my profit and revenue? I think we're just fortunate in the company to have people who have worked at [inaudible 00:27:21] or banking, such as myself, or retail, such as a colleague of mine, Patricia. So all coming in from various business fields, but we're all coming together in the interception, which is data. We're all going to use data to improve businesses, so we dispense that advice.
Then on the third leg, we also do training, so especially in the Philippines where we have to admit skills are still in short supply, so there's never a shortage of people who want to do training, so we do that as well, whether it's in-house training or public training. We're not really marketing ourselves as a training company though, but it is a good source of leads, so that's another ... want to get into consulting. For those wanting to start analytics service companies, do consider training, which can be very complimentary. You can test ideas and products in the training classes. Of course apart from booking a little revenue as a trainer, you can use that as a rich source of leads. Normally the ones who would sign up for our training classes incidentally work for companies who do need analytics services, so it's been a very, very helpful and successful for us in the past 24 months, finding customers attending these training classes.
Kirill Eremenko: That's very cool. Thank you so much for sharing and diving into the description, so I'm just going to recap on that. Especially I think it will be useful for those who are considering starting a business, or maybe like somebody listening to this podcast might not be considering it now, but maybe one day you'll come back to it, and you can re-listen to this bit.
So vertical one is data and engineering, where you do the whole suite from data sourcing to BI, ML, analytics, and you make it friendly for the business user, which as you mentioned, is a critical point. Then you got vertical two, which is the consulting side of things, and you more step away from the technology, but you talk about the use cases for the specific company that you're working for, and make the approach tailored for them, so they realize what they can get, what value they can get out of analytics.
Vertical three is the training component where you have in-house and public training, and those are great, rich sources of leads for you and your business, because people who need training, incidentally, they're most likely working for companies that might need analytics services. It's a really good set up. I can see how you have lots of synergies between the verticals.
Dominic Ligot: Yup, yup, and it's also good to attract talent that way, so normally if you can't find clients in these training classes, you will find a future freelancer or a future collaborator, because they suddenly quote unquote, "See the light," and say, "Hey, I've been looking for this all my life, and now you've showed me how I can become more productive."
In fact, a couple of the guys who are working with us now started out as students, and they since done a career shift. That's kind of like a lesser, I think less taxing way to get into the industry is rather than go full on and start your own company, maybe find a start up that you can apprentice with, or do some freelance gigs with. I think over time, there will be more and more companies such as ours, who will be on the lookout for talent, and training is a great way to find them.
Kirill Eremenko: Gotcha, gotcha. And speaking of talents, I have a bit of a more of a business question for you. You mentioned you're 10 people right now, and what I was wondering is are you planning to grow the business? I've seen two types of ways consulting, analytics consulting firms, can develop. One way is when you keep growing, and you grow into a larger, more mature analytics business where people are trained in the different components of analytics, for instance, in the different parts of the verticals that you described, and you have specific people doing specific roles.
On the other hand, there are businesses who choose to stay smaller, more boutique analytics consulting firms, but they train up their staff to be like Swiss Army knives of data science, and they can do almost anything. They can still be competitive with 10 people, and because it's such a small firm, they don't have the large overheads, and yet they can still charge large fees for their services. So there's kind of like two ways that I've seen analytics firms develop. What is your plan for your business, if you don't mind sharing, of course?
Dominic Ligot: Yeah, yeah, so that's a great point, and I totally agree. I don't know if this is going to be counterintuitive, but we're going to be more of the latter, for the most part. One of the things that so far over the past 24 months worked to our advantage is we're quite fast in delivering outcomes for clients, and that's why they stick to us.
If we're going to grow, and most of our manpower comes from freelancing, so it's probably not going to grow that much from the current size, in terms of the actual hands-on work. You hit it on the nail when you say the goal is to make Swiss Army knives, so like jack-of-all-trades. Like if I'm going to talk about myself, I did start from say the business side of things. I got into the IT side, learned the engineering, and then in my past life I was in banking. That's where I picked up some of the statistical knowledge on the data science.
So I'm a little bit of a jack-of-all-trades, and that's also how I found the people I collaborate with. On the other hand, we are conscious that there are some parts of our verticals that are growing really quickly relative to market demand, and that might deserve a second look. For example, the training that I mentioned earlier, there's a huge demand on the ground for us to do training, and now it's actually coming to a point where the training's getting in the way of actually doing the job, or doing the rest of the work.
Some of us enjoy doing the work more than teaching it, so that is a serious consideration to us to say as early as 2019, 2020, do we spin off, say a proper analytics training center, a proper school for analysts? Or do we even go further than that and become more of an analytics recruitment center, where we come in in a sausage factory, give you the training, and then place you in a job, all of which could be lucrative, at least in the near term?
Those are serious considerations, but my default position would be we're doing well at the moment, being a good vendor, we're doing great work, we're actually looking to start creating a few products that our clients can actually start subscribing to, so you get a little bit of passive revenue without doing extra work. Then in the medium term think about spinning off more proper division, let's say for training, which could be a good play in this kind of market environment.
Kirill Eremenko: Okay, gotcha. Well, thank you very much for sharing. I hope none of your competitors hear this, because you're sharing everything on your strategy. I'm sure everybody appreciates it [inaudible 00:34:46].
Dominic Ligot: Yeah, we talk about competitors, and again, this is hopefully it doesn't end up shooting business in the foot, but I'll tell you why it won't, because right now, I don't know how many of your listeners will be able to relate to this. This data analytics industry is really still quite silo, so we mentioned data engineering for example. Even that isn't really properly defined, so you've got some IT people who know how to extract data, and maybe you have a few DBAs who know how to put it in a database.
Then you have a few analysts who kind of know how to get it out of the database, put it in your Python notebook, and come up with some visualizations. Then you have another application developer who will get the output of that, turn it into an app. So you need those four people to really cooperate, and the irony is, you have IT vendors, you have database vendors, you have analytics vendors, and you have application developers, but they all only know their little piece of the pie, so there's really, really opportunity in stitching these things together, being more of a full service or say generalist type of vendor.
I mean, it won't fit everyone, but there is opportunities in stitching together several things. I think the industry will see more of that, because you don't want to be paying four different people to do what one vendor or two vendors can do, or do really, really well. As I said, we don't really play the enterprise space, so the clientele we attract aren't also the type who would be hiring like five or six different vendors from mega companies. They prefer a one-stop-shop.
I think that's an opportunity, especially for people who are starting out. I think maybe easily nine out of 10 people I meet who are starting out data scientists, kind of focus more on the analysis. While that's very good, it's very, very rich field to get into it, there's a lot of things to do, but don't ignore the other parts of the value chain. While you're studying your R and your Python, or maybe your data visualizations, your Tableau, don't forget the back end, because that's where the data's going to come from.
Normally when you get into a job, even if you're not starting a company, you're going to start out as an employee. You're going to have to do that anyway. You're going to have to run a few queries, get data from someplace. The company would appreciate if you could do both rather than have to rely on the IT department to do that.
So that's kind of an open trade secret that for some reason, people are collectively ignoring, or at least maybe in my end. Maybe everywhere else it's already developing, but rather than keep it to myself, I think we will all benefit if people become more and more multidisciplinary as a result.
Kirill Eremenko: Yeah, yeah, totally, totally agree with you. You gave me the story as well that the analytics industry is also not mature at all, as opposed to the accounting industry, or like some finance, areas of finance. There's lots of room for many companies. I think it's very admirable that what you're doing by sharing this information, because ultimately, instead of making even competitors, like instead of competing with companies, companies can create alliances and work together.
Dominic Ligot: Absolutely, absolutely. Just as an example, another one, and we can talk about this more later, I said we're not a proper training company yet, but we're trying an experiment in September and October. We're going to run a few niche classes, and the target is really not data scientists or data engineers, as such. The target would be business decision-makers, and maybe business analysts, and run them through what we would call a Masterclass, where we can take them through the entire value chain.
"Hey look, this is where you get the data. Hey look, this is how you store it. Hey look, this is how you analyze it." But rather than focusing on what everybody seems to be doing in training, which is teach code here, teach software there, of course that's important, but no one's actually out there teaching the business decision-makers, so just exactly why do you need the data, or how would you use these types of reports?
That's another niche that's waiting to be filled in terms of now you've got a rich source of data, you've got a lot of tools at your disposal. Maybe you've built your analytics team, but then the gap is what are they going to do? I mean, they don't speak the same language as the business, or vice versa, the business people don't speak enough of the data language to translate their objectives into analytic models and strategies.
Then that's it, that's a lot of sunk investment right there. As a smaller niche to that, just to put it on the table, now that we're getting into more automated decision-making, more algorithms, there's a looming need for what I would call data ethics professionals, so if you think about the stuff that recently happened with Cambridge Analytica and Facebook, on the one hand, or a couple of months ago, the self-driving car ran over someone in Florida, and that was purely on the basis of the failure of some object detection process.
So now people are getting hurt and they're dying because of data, and no one actually seems to be stepping up and saying, "Hey look, there should be this code of conduct or ethical standards when you use data, in the same way we have similar things for medicine or law." When you get into a more mature field, there is an ethical line that needs to be drawn and how these things are being used.
The only thing people seem to talk about now is privacy, and that's the tip of the iceberg when it comes to ethics. That's another field that I wish maybe I had more head space. Setting up a data ethics consultancy would be probably the thing for 2020 and beyond, so you know, it's just an early shout out for people who are coming from let's say the legal profession, or the ethics profession. Data is out there waiting for you if you want to do something.
Kirill Eremenko: Wow, fantastic. Thank you for those two use cases of data and training. You mentioned the executives or the business decision-makers training, and ethics. I understand this whole ethics side of things, and think you describing quite a bit of detail has got a lot of opportunity. But I'd like to talk a bit more about this business decision-maker training. How did you come up with that idea?
It's interesting how we haven't spoken before, but we're thinking in the same direction, because that's exactly what we're focusing on right now. We've also identified this as a niche, and we're thinking, "How can we help executives and business decision-makers better understand data and better use it to their advantage to help grow the businesses?" How did you come up with that idea?
Dominic Ligot: I think a lot of it is inspired by I guess my own adventures back in the day when I was working in banking. I spent 14 years in banking before I went into IT, and I was a business decision-maker. Through numerous frustrations, because I couldn't get an analyst to cough up the report that I wanted, I kind of ended up doing it myself or getting my own people to do it.
Then on the other hand, numerous struggles with IT to source the right information, because if you're a decision-maker, you need information, and the information doesn't come easily, especially if you're in a company that's not quite mature. That was the primary inspiration. There's probably tens of thousands of people exactly going through the same challenges that I did, and there's nothing out there that's helping them, so that's in the first instance. Maybe if we cough up something people would be interested.
The other thing, I guess from a broader perspective, that I'm usually pretty conscious of ... let's say, I would call it changes in let's say generational habits, so everyone calls ... everyone groups people into like 20 year buckets, right? We have these Baby Boomers for the first 20 years after the war, then the Gen X, and then now we've got the Millennials, and now you have Gen Y and Gen Z. So all of them have very, I would say as a general group, have different habits.
One thing that has made a big change now, particularly as we approach 2020, is many companies are being run by Gen Xers and Millennials, and the big difference between these guys, including us and our parents and grandparents, is we grew up in a very digital environment. We played computer games, we ... internet, and we kind of want to manage businesses that way, you know? The biggest inspiration for analytics is I think computer games. You want a score card, you want an indicator of how many customers you tap. Everyone responds to that quite naturally.
Intuitively, you know what data is supposed to be used for, but in terms of the availability of proper training out there, like, "Hey look, if you respond with a score card, what does a score card look like for your business?" For example. Or if you like using apps like Google Maps, or Waze, like Waze is pretty popular here, and you use that to get around, what's the app that you need to help you navigate your business strategies? Do you have an equivalent of a Waze or a Google Maps for your business?
That takes a lot of not just number crunching, but a lot of insight. You need people to be guided to think about data in a certain way, and whether you're in HR, or marketing, or an operations job, whether you're in finance or [inaudible 00:44:43] the needs are very, very similar. You want to make sure your business is viable. You want to be sure you make money. It's very rare that you can find opportunities to link data and that together, so that was kind of the background.
I said, "Okay, why don't we start listing down what are the typical use cases for say marketing?" So marketers want to acquire customers, or they want to retain existing customers, or they want to understand why customers are complaining or about to leave, so these are very normal things marketers do. But then guess what? Now that we're in additional era, all of this have an equivalent data point, and analysts know about those data points, the engineers know about these data sources, but it's in stitching it together, that's the crucial mix.
So that was the inspiration for the Masterclass, and yeah, hopefully it works out. The initial response has been very positive, but again, you run into the usual challenges of since it's never been done before, or it's very rare, no one knows how much to pay for it, or whether it's worth paying for. That's the proverbial kind of first mover issue that needs to address. But yeah, I think it's the way people should be thinking about data moving forward.
Kirill Eremenko: Yeah, yeah. Another challenge I find with this type of masterclass is as you say, because it's something so new, business decision-makers don't really know how to pitch it to the board of directors, or to their managers. Not ultimately you're going to get the CEOs, they just might be like the CTO, or it might be just like a high-level manager.
They need to include in their budget, right? So they don't really know how to pitch it to their manager to say, "Hey look, I need this training because it's going to benefit the business." Then their default fall back is thinking that it is an out-of-pocket expense for them, and because ultimately you cannot run this as a ... the same way as you run a training class, you cannot get like 100 people in the room, you can only do it very specifically-
Dominic Ligot: Very small.
Kirill Eremenko: Yeah, yeah. You want like 10 people in the room max, or 12, I don't know. Because of that, the price is going to be high, and then they got this dilemma, then on one hand, they know is business value, they don't know how to pitch it to their boss. On the other hand, it's very expensive, so they can't really pay out-of-pocket, and they're like, "You know what? I'm just going to probably pass on this opportunity," when it's ultimately, it's the thing that's going to change so much, because if you change what's happening at the top, the whole business changes.
Dominic Ligot: Yeah, yeah. Pitch, it's almost like it's not just a business transformation issue, it's a cultural transformation issue, if you're not used to thinking about data training, or analytics training as a business expense. As I said, this probably will probably end up by default in the IT department or the CIO's purview.
You do meet on a rare instance sometimes that it is the IT department that's encouraging the business to join them, and that's usually a good peg they find, or you have the newer type of executive, like a Chief Digital Officer, or a Chief Data Officer who kind of sits in between IT and the business, and normally it's their initiative to get into this, but that's kind of a rare thing.
On the other hand, just as another tip, what I've seen work really well is if you land or find a company that are hitting the proverbial brick wall in terms of their growth. They used to be a small company and they've hit the medium sized level, and they're still running it like a mom-and-pop shop, and now they're suffering.
Or the other way around, like you have a medium sized company, and they're about to enterprise territory, but they're still kind of doing everything manually and now they're suffering. So there's suddenly a pain point that they can't address, and that usually gets an audience with the decision-maker, and you say, "You know what your pain point is, is that you're still doing it the way a small company does it." That's where data and analytics can come in and sort it out.
You kind of see it as, "Okay, I'm not really sure if what you're telling me is true, because I've never heard of it before, but it's worth a little experiment. Okay, maybe I'll send five people or six people," and then you take it from there.
It's a maturity thing. Over time it will become normal. If you can imagine maybe 15 years back, people were thinking about e-commerce and the internet pretty much the same way, like, "Hey, I need a website," or, "What kind of digital marketing do I need?" Even back in the day, people refused to acknowledge that digital marketing was part of marketing.
So, "Yeah, we're a marketing department, but digital marketing's that guy, and he's part of the IT department." It's the same. I mean, now it's a given. If you're not online, it's marketing suicide. Chief Marketing Officers need to have a digital strategy. It's just this first hump that we all need to get through, but yes, it's fascinating that you're getting through the same challenges we are, and yeah, maybe we need to have more discussions like this to understand how we can do it better.
Kirill Eremenko: For sure, for sure. Well, that's been very exciting, and we're slowly coming to ... like it's crazy how time flies. It's already come to the end of this podcast. I have an interesting question for you that I would like to get your opinion on. For what you've seen like 20 years in the industry, and now you've moved to consulting, doing your own consulting in the space of analytics, and growing a team, and helping other businesses, where do you think the field of data science is going, and what should our listeners look into to prepare for the future?
Dominic Ligot: Well, I think from what I'm seeing in the local market, and I think this kind of mirrors what's happening across the world in various degrees. Maybe three big trends I'm seeing. The first one is democratization of knowledge and skills. Back in the day, when data science wasn't even a term, it was very hard.
When I say, "back in the days," like as late as the '90s, very hard to find information about analytics. You had to find special books, and you had to find special people, and you're usually stuck in statistics departments and computer science departments, and they don't talk to each other.
Now we're seeing every major university coming up with some sort of a data science course. I think that's more good than bad, because the one thing everyone still struggles with is what is the proper definition? I'd rather not get into that debate anymore. It's more about, "Hey, you know what? Learn as much as you can, because the market's waiting for you."
And of course the internet has been very helpful, the rise of open source trend, and everyone now can pretty much learn Python and R, and all these open source software on their own, watch a few videos, listen to podcasts like these. So there's never been a time in history for knowledge has been more democratic, but adoption has been slow.
That's the second trend that I'm seeing. I think after 2008, so as a banker, that was a pivotal moment for me is the financial crisis. That financial crises have a habit of knocking out businesses that aren't robust, are inefficient, and that's given rise to a more conscious need to, "Okay, how do I end the competition? How do I get ahead?" Margins are getting slimmer every day, and regulations are getting tighter.
The need for that new thing, that new Holy Grail to get ahead of businesses, data is one of them. Of course you have other big, big, big stuff, like the usual stuff like blockchain and all these other trends. So I would analytics falls smack dab in the middle of that. Before it used to be niche, like it's a luxury. Now you have companies, the most expensive companies in the world, like the Googles and the Facebooks. These are all data companies. It's foolish for you to ignore it.
In a country like the Philippines, which is pretty protected, more and more industries are getting opened up to liberalization and market competition. We only need to look at what happened to Uber, for example, and Grab for Asia, and how that's messed up the taxi industry, and see how getting a bit of data and analytics into your business model can really, really be destructive. So that's the second trend. You're going to see more of that moving forward.
I think the third one is interesting, because there's a subset of analytics, and for me at least it's a subset, which is the whole area of deep learning and artificial intelligence. It's still for the most part, I mean, if you listen to the media, it's still in science fiction territory. Everyone's worried about the rise of the machines, and the terminators, and all that.
I think that's an area which is interesting to watch, because the more you think about it, the more intelligent algorithms start to permeate processes in the workplace. I don't think it's necessarily machines rising up against the humans, but it's more about how do humans work together with machines better? It's not going to be Kasparov versus IBM, Deep Blue anymore, it's about how do I get a chess algorithm to beat a normal chess player?
That's going to be interesting, because when machines become independent, you shouldn't be worried about how they're going to make your life horrible, what's exciting is to see how they're going to make your life more efficient and better. When cars start driving themselves, imagine how more efficient that will make transportation, for example. There's going to be an analog everywhere else you go, and AI, machine learning, deep learning, these are not easy things to do, and that means there's going to be a massive demand for people who understand not just the technology, but the maths and the sciences behind it.
Again, there's never been a better time to get into the nerdy stuff, like computer science and math. That's great. I mean, Gen Z and Gen Alpha, there's a very good chance most of them are going to be a data worker of some sort, just like once upon a time there was one computer operator in a workforce of 100, and then now everyone has a laptop. Now you've got maybe a couple of people who know data in a workforce don't. It won't be very long before everyone's a data worker at some level, so there's a lot of new jobs that can come out of that.
Kirill Eremenko: Fantastic. Thank you so much for such a detailed overview, very insightful. I'm just going to recap. Three big trends that you're seeing, so first is democratization of skills. It's never been easier to learn things, especially with online. Then second trend was the proliferation of data science, data science is becoming more commonplace. An example such as even Uber showing how disruptive it can be, is those things are pushing businesses to not see data science as just like a ... something like a nice toy to play around with, but something that is going to become part of their operations, like an integral part of their business.
The third trend is machines working with humans, and that is concerning more AI, machine learning, deep learning. The complex things or nerdy things, or the things that used to be considered just nerdy, are now becoming more and more as well commonplace, and they're going to be helping us make our lives better, so it's a good time to jump onto this trends.
Dominic Ligot: Yeah.
Kirill Eremenko: So thank you so much, Dominic, for sharing all those insights. I'm sure lots of people learned a ton. I personally learned a ton today from you. If anybody would like to follow you or learn more about you, things you share, follow your career, maybe you can get in touch, what are some of the best places and ways to contact you?
Dominic Ligot: Well, I'm really only on LinkedIn on a personal basis, so just hit me up on LinkedIn. You can look for Dominic Ligot, or dot Ligot on LinkedIn. My company's on Facebook though. I don't have a Facebook account, because I'm such a privacy nut, but the company's there. I mean, if you know enough about data, it spooks you out too much, so interestingly.
But yeah, you can search for us on Facebook, just search for Cirrolytix or Cirrolytix Research Services at C-I double R, O-L-Y-T-I-X. Then we also have a website. You can find us at Cirrolytix.com. If you're based in Asia or in the Philippines, you might be interested to hear about our Masterclass, so the URL is upskill.ph, so U-P-S-K-I double L dot ph. Or you could also look for our business analytics masterclass on Google, and I'm sure it's one of the things that will pop up, so yeah. Looking forward to linking up with you guys.
Kirill Eremenko: All right, and thank you very much, and we'll definitely share all those links in the show notes. I just have one last question for you today. What is a book that you can share with our listeners to help them in their careers?
Dominic Ligot: Okay, so there are actually two. One is an older one and one is a newer one. The older book that I keep defaulting back to, and this is not a technical book, because there are too many of those already. There's a book called Competing on Analytics by Tom Davenport. That for me has just in the classic bible for me in terms of what differentiates a company who uses analytics for not just as a toy, but for competitive advantage, versus the ones that don't. So yeah. I'm sure if you read more of Davenport's books, he talks about very similar things moving forward, so that's one.
The other one is more on the philosophical side. There's a book called Life 3.0 by Max Tegmark, and he talks about all the hypotheses related to AI from the really crazy ones where the AI enslaves us, to the more I would say realistic ones, where we kind of merge with machines, eventually, and that kind of gives us the next step in the evolution.
Now, I like that book, because not only does it spark the imagination, but it also gives you some practical grounding to look forward to, like why are we all studying this? Why is this a big deal? I think the secret is, it's a major part of human existence now. Data is us, and the digital and the real world are blending together very, very quickly. The future belongs to those who understand data very well. It's the new real world, technical.
Kirill Eremenko: Totally agree, totally agree. There's lots of movies to portray that, that came out recently. Just to recap the books, Complete Analytics by Tom Davenport. By the way, for our listeners, Tom Davenport is the person who together with D.J. Patil wrote that article that proclaimed data science is a sexiest job of the 21st century.
Dominic Ligot: [inaudible 01:00:03].
Kirill Eremenko: Yeah. The second book was Life 3.0 by Max Tegmark. So thank you so much, Doc, for coming on the show today. Once again, really appreciate you spending the time taking out of your busy schedule to share all these insights.
Dominic Ligot: Thanks for having me.
Kirill Eremenko: So there you have it, that was Dominic Ligot. I hope you enjoyed today's episode. I personally enjoyed it a lot, and also learned a ton. Probably my favorite part of today's show was just the variety of business tips that Dominic was supplying, and the fact that despite the temptation, we didn't switch to talking about the technical aspects.
I know that probably a lot of you are thinking that it would have been nice to talk about the technical [inaudible 01:00:54] but we have lots of podcasts to choose from in that space. Here I think the value and the advice that Dominic was sharing in the space of actually building a consulting business in the space of analytics was extremely valuable.
On that note, if you'd like to get the show notes as usual, you can get them at www.superdatascience.com/183. There you will also find the URL for Dominic's LinkedIn. Make sure to connect and get in touch, and especially if you're in Southeast Asia or in the Philippines, then reach out to Dominic and maybe attend one of his training sessions. Maybe he can help you with some consulting work or maybe you can just exchange some information about what's going on in the space of analytics.
On the other hand, if you know somebody who is in that region and who might benefit from connecting with Dominic, then be the connector and connect those two people. I'm sure they'll say thank you to you at the end of it. On that note, I hope you enjoyed today's episode as much as I did. Can't wait to hear you and see you back here next time. Until then, happy analyzing.