Intro
Hi and welcome to today's SME Business podcast, your host, Mark, will be joining you to interview a founder of an SME business each week, highlighting lessons learned and revealing insights. Listen and learn each week on how to get and stay ahead.
Mark
Hi Listeners, thank you for tuning in for the next episode of the SME business podcast. Today, we're joined by Colleen Tato from Starwars Data. She has spent more than 15 years in a data engineering and analytic space. She did a PhD in astrophysics with a dissertation on the potential feedback mechanism between the intergalactic medium and dwarf starburst galaxies. And Starburst Data is a data analytics company developing a SQL query engine. Welcome to the show.
Colleen
Thank you. Happy to be here.
Mark
Yes, so to give a bit of intro to people listening to the podcast, of course, I know you more already, but can you tell listeners a bit more about yourself and how your journey went?
Colleen
Yeah, absolutely. Thanks, Mark. So I am calling Colleen Tartow, a director of engineering here at Starburst Data in Boston. And I do think I've had a bit of a roundabout journey, but maybe one that's not so uncommon. I began my career in astrophysics, and I'm very impressed that you got the title of my dissertation; that's a first. I'm impressed by that. As you said, I was studying the intergalactic medium, which is the space between galaxies.
Colleen
And it's interesting because at first, it looks empty. Still, it turns out that there's a lot of stuff in between galaxies, and it comes from the galaxies themselves. So the question is, how did all of that stuff get to the intergalactic medium and how does that affect our understanding of the universe? And so I spent several years doing that and I had much fun with it. But ultimately, the work itself was focused on taking giant data sets and slicing and dicing the data, which probably sounds pretty familiar to many people.
Colleen
But I was spending most of my days locked in front of a computer and doing fairly esoteric work. And, it's a tough, tough industry to be in academia. So I decided to make that jump from academia into the "real world", as my parents called it. And I started making my official career. I just kind of fell into it. But I started working in Big Data, and I worked in Big Data for a while.
Colleen
And then I started to think about, well, what are people doing with this big data that we're helping them get a handle on? And so that sort of turned into a career in analytics and data science. So that's sort of always been my passion is data and data science. And that's really what I had been doing all along anyway, which is kind of funny. So I started as a consultant helping enterprise companies build out big data and analytics solutions. Then over the past five years, I focused more and more on building data and analytics organizations within small and midsize companies.
Colleen
And I sort of pivoted a little bit and came to Starburst data to help build software for data and analytics organizations inside companies of all sizes. So at Starburst, we are building really exciting software that I truly believe will revolutionize the way people think about data and how they get business value from data. So that's sort of where we are today.
Mark
Yeah. So as I was going to say, and I like the astrophysics background, it's a cool idea. Dear listeners, I'm sitting here having a massive imposter syndrome. I have a daughter that wants to be an astronaut and has a telescope. And I like to think that at one point, I will understand dark matter. And I sometimes pretend I star in Stargate Atlantis and go through a stargate. Those are just, I guess, fantasies for most of us, I guess.
Mark
So I'm not going to pretend I know this kind of thing. So maybe at one point, my daughter can teach me, but yeah, amazing. I'm not going to like to ask technical questions about astrophysics because I will not be able to. You can go in-depth, like, I was good in chemistry, that kind of thing. Yeah. But physics wasn't my strong point. And funnily enough, economics I was better in, but yeah. Numbers were kind of boring.
Colleen
It's all Data.
Mark
Yes. In the end, this all data, and let's be honest, we've all been in a situation where we did way too much with Excel then we should have been doing. So yeah, it's great there are more solutions out there that allow people to manipulate and analyze their data because that's one of the things where your advantage just lies for organizations where you do more with the data that you have, and then that gives you more insights.
Mark
So yeah, it's good to see, and that's good to see as well, organizations that are not coming from the standard Silicon Valleys of the world. So it's good to see the support that way. I want to bring up something as well. I know you're a huge diversity and inclusion advocate, and I was wondering as well, how can small and medium enterprises listening to this podcast do their best to advance that agenda?
Colleen
Oh, that's a great question. And it's one that I think about a lot. And I, I give quite a few talks about things like this because there's a lot written about it. And I think companies struggle with where to start. So to start, I think it's important to understand the difference between diversity and inclusion. Diversity is having people from a variety of different backgrounds, whether that's gender, age, race, education, economic status, et cetera.
Colleen
And inclusion is having an environment where folks from any background can thrive. So you can be diverse but not have an inclusive environment and vice versa. So the key is to have both of these things, have people from various backgrounds, and have an environment where those people can thrive. So they often do go hand in hand, but they are different. So I think that sort of step one and then step two is really how do you become more diverse and how do you have an inclusive company?
Colleen
And I think part of it is making a case for it. And the case has been made. Right. Like you could read any one of a bazillion studies out there that diverse and inclusive companies are more innovative. There's something like 70 per cent more likely to capture a new market if you're a more diverse company. You can also there's data out there about once the number of women in management rises above 20 per cent, the revenue rises by X per cent too. It's very clear that the more diverse you are, especially gender diverse, the more successful your company would be.
Colleen
So it can be overwhelming, though, because you can make that business case. But it's like, well, where do you start? So for a smaller business, I would recommend picking a few aspects of both diversity and inclusion and focusing your efforts on starting. From a diversity perspective, a. Action is to source candidates from a broad range of places; companies, especially small companies, tend to rely solely on referrals and their networks.
Colleen
And it's tempting because referrals are great. There are people; it's like a known quantity you know that they're going to work out well, and you have that history with that person. But it's been shown time and again that these are incredibly limited demographically, and you end up with a homogenous workforce. So as tempting as it is to hire referrals, I think you need to step back and limit the number of referrals you're bringing in and try to source from more diverse places.
Colleen
And so that means you need to do things like sponsored diversity, diversity-focused organizations, look outside the box for where you're thinking of sourcing from. And then, on the flip side, from an inclusion standpoint, assuming you have a diverse workforce, you want to make sure that you have an inclusive environment. And that means things doing things like promoting fairly, ensuring that your promotions aren't tied to specific people, but that you have a very clear, articulated career ladder within your organization, providing training across the board and then making sure that your company policies aren't biased.
Colleen
So yeah, I think picking a few specific actions that your company can take and then getting buy-in and taking those actions is important.
Mark
Yeah, that sounds a good thing; depending on the size of businesses, it's easy to try to pick up too much and try to do too much. And then you end up not doing it well at all. Yeah. So, it's good to pick a few things and then improve from there. And then when you bring new talent in, and those can be like, if you will, the champions to help you go further with that agenda.
Mark
And like I said, in the end, it is good for business because if we have different opinions and such, it's only good because you don't want to have a situation where you have meetings, and everybody just nods and say yes. And you end up potentially going along with something that is not relevant anymore because you don't have the right voices in your team.
Colleen
Absolutely, yeah,
Mark
so, can you tell me a bit more about Starburst Data and how it started?
Colleen
Yeah, absolutely. So I've been at Starburst for almost a year now, which is very exciting. It's based here in Boston. We also have an office in Warsaw, Poland, and we have an office; I think we will have an office in California, but we are everywhere. And we're a data platform company focused on really shortening the distance between data at its source and the business value for our customers and allowing them to do analytics anywhere. Our flagship product, Starburst Enterprise, is based on the Open Source Project Trino created by folks who work here at Starburst.
Colleen
Trina was created at Facebook about ten years ago. It was originally called Presto, and it was created in order to solve the data and scalability challenges that companies like Facebook see a lot. And that's things like the fact that you have to move the data away from the source to analyze it. And that inherently introduces a lot of complexity and complications. And so Facebook, they were like, hey, let's not move the data in order to do analytics.
Colleen
And so they created this product, and now called Trino, it's gained viral adoption. And then, Starburst Enterprise was created as the enterprise version of the Open Source Project. It was first launched in twenty-eighteen. And the ubiquity of cloud and Kubernetes, we can run virtually any environment, we can run on-premises, we can run in a data centre, but we can also run on any cloud environment. And so the idea is that you connect directly to your data at the source, and you can just use SQL to talk to your data, just the lingua franca of data.
Colleen
So everybody is equal. And now we're breaking into the managed services side of things by offering Starburst Galaxy, which is essentially Starburst enterprise as a service and then some. And so, we host the infrastructure. And that's actually what my team is working hard on delivering. And we are in an early access phase now, and we're seeing great feedback from our beta customers who are very excited about that.
Mark
Yes, yes. And I was going to say, I read about the announcements, so I would say it's such a big deal. And I can see like especially the managed bit of it makes it easier for customers that might not have the technical bandwidth to manage such a thing themselves. And the data that organizations are collecting is getting bigger and bigger.
Mark
If you go to very large organizations, they're collecting terabytes of data per day. It becomes harder and harder to analyze because no such thing anymore, were back in the day, maybe you're collecting customer records. These are all the invoices for this customer. And now you have so many dimensions, if you will, of data. We give you, like, hey, what's up.
Mark
I can not only see the invoices, but I can be like, hey, this customer's always ordering parts that are if you're talking about steel like over 60 centimetres long or something. And more people are doing that. So maybe it's that something of a decision point for us to say, hey, we need to start producing longer widgets because that's the advantage of data.
Mark
You can see trends from that. And then it can help companies make more money, that kind of thing.
Colleen
Yeah. And I think there's also, there's sort of the older generation, the legacy data that you might have and like larger enterprises often have that legacy data as well as new data. And the idea is that you want to both analyze the legacy data and analyze the new data. And as both of these things might change over time and then combine them, and you might have data in the data centre, you might have data from some cloud-based services that you offer.
Colleen
And so the idea is Gartner calls this X analytics the cross-section of that old data and the new data and bringing it all together. And if you have to build a pipeline for each data set that you want to bring in, you can't just say, well, it's a 15 second SQL query. It might be a month plus 15 seconds because you're trying to actually build a pipeline and then run the SQL query.
Colleen
Right. So the idea is that if you don't need to build a pipeline, you can just get to the data at the source with SQL. Then it doesn't matter where your data lives, and you can connect to data anywhere. And so that's really what Starburst is all about, that it is analytics anywhere.
Mark
Yes. And if that's not already a company sub slogan, then I would suggest putting that in.
That actually is our slogan. Yeah. Analytics anywhere is our tagline.
You could almost do advertisements and then do car apps or something. Yeah. iOS, carplay ready one for a few sales managers, and be like, here's your dashboard these are the customers you need to talk to, or you need to ask these questions.
Mark
That's an interesting thing for me. Sentiment analysis, where you could be like, hey, can we look at the data and figure out what the customer's requirements currently are? And I think that that's going to be interesting more A.I. and such and then more data, I guess it is a good thing with data as well. I just think about like back in the day, like every organization pretty much was building a Hadoop data lake. Then they're collecting all the data, and then they're like, OK, but yeah, now we need to do something about it. So it's good to know that Starburst Data, in that sense, can help with legacy data because I know plenty of organizations that will have a data lake somewhere. They collected tons and tons of data, but nobody's willing to admit that they haven't been able to create use cases on top of it.
Mark
So I think that that's going to be interesting as well. And, and I figure, in the medical space as well, it's going to be interesting, like things like homomorphic encryption or whatever without getting too technical where we can anonymize data so we can do data set analysis over much larger data sets to figure out like for instance, like new symptoms and such and be proactively being able to warn people, be like, hey, you're coming to this hospital, you live in this region, you have these, kind of, symptoms.
Mark
So this is probably something you need to check with your G.P., or your M.D., depending on which country you live in. So that's going to be interesting, like, but yeah, I think privacy and such, it has to be top of mind because there is a thin line between getting their functionality and then people feeling intruded on, that kind of thing.
Mark
Yeah. But that's going to be interesting. I think in that space where I think over time, it will surface like early indicator symptoms for illnesses. Maybe right now, we are not now diagnosing this until a much later stage. I think that's going to be the exciting bit. And then we'll be like, hey, what's all these people, for instance, have this specific illness?
Mark
We all discovered that they, for instance, have a specific marker. In their blood, Because we saw it in one person and we checked the other people, and they have the same marker, that kind of thing, so we can be like, proactively, hey, you need to go to G.P. or an M.D. and, to get a blood sample taken because if we do it now, then we can still do something with it.
Mark
So that's my hope, so maybe one day. Any startup working in that space, just call me. I will happily provide some free advice. And I have some questions that I always ask. Yeah. Of course. Secretly I hope that the government officials worldwide start listening to this podcast if they don't already do that. And if you're not, then why are you not doing this?
Mark
I'm Dutch, and I just say how it is. So one of the things is that questions like: What could governments around the world do better to help small businesses grow?
Colleen
Oh, yeah. I mean, I'm no economist, but having spent time as an employee in companies of all sizes of all over the world, I will say that especially in the U.S., it does seem that like the government needs to close the gap between the lip service that they give small or medium-sized businesses. And the actual tax breaks that go along with that. There's this huge dichotomy in the political machine between what's promised on the campaign trail and what's delivered via legislation.
Colleen
So I do think that at least here in the U.S., there needs to be more representation within the government itself for the specific challenges that small business owners face because small businesses lift economies up. And in parallel, you can't say you support small businesses but then give huge tax breaks to giant corporations. You need to be able to support both of these things in any modern economy.
Mark
I'm willing to go on record and say, Amazon, I'll do that. Jeff Bezos, if you don't like me, that's fine.
Colleen
I was going to say Wal-Mart.
Mark
Yeah, yeah, yeah. Anyway, just in general, like large corporations. Yeah.
Colleen
But I do think anybody on the list of the top richest people should probably not be getting huge tax breaks in the same way. But that said, I also think that as you think of how we make small businesses grow? Like how does a government determine what's a small business? Right. Like when the U.S. had all these applications for small business loans when the pandemic hit, they had the first relief act.
Colleen
It's like, OK, well, functionally, how do you give checks to the right people? And that's all data. And it's like make sure that you have a data solution where you can, like, pivot and be like, OK, give me a list of the small businesses. How do we determine small businesses? Great. Here they are. Should Wal-Mart be on that list? No. Is it? Yes. And so making sure that you have accurate data, I think, is something the concrete action governments could take to ensure that they've got the right data about different businesses.
Mark
Yes. Yes. And it's like it's one of those things where I always like remark it as well as like the Fortune 100. A lot of those companies don't end existing in the next 20, 30, 40 years. They disappear. And I am talking about Amazon again. I think even Jeff Bezos, in his last letter to stockholders, I think. He said they should like bet on Amazon not existing anymore because he says companies rise and die.
Mark
So my point is, in general, is this as well. It's in the government's best interest to help support small businesses because, in those small businesses, there will be the new large corporations in there that they need for the future Fortune 100 and or Nasdaq in New York. I've Nasdaq and such because back in the day, nobody said that Blockbuster would go away.
Mark
And yet, there's no more blockbuster. Yeah. I remember like back in the day; I remember people saying, like why would people want to be on the Internet? To be fair, it wasn't called the Internet. But I just remember paying for dial-up by working in the supermarket to pay for my dial-up cost. I remember having a home page. It was called out on the home page of the Internet provider's website because I was one of 12 people who had a website, which is like if you think about it now, it's so crazy.
Mark
In the Netherlands, you had this thing: I think Yahoo had something similar, but it's called Start pagina, which is like a start page, roughly speaking. So you had a start page for, like, different subjects. So you had like start page for nurseries or whatever. But back in the day, it's funny.
Mark
Suppose you say now you like people were laughing. In that case, you go like you, for instance, like there were so few websites that all the nurseries in the whole country in the Netherlands, for instance, fitted on one single page, which like crazy if you think about it. I mean, you go like really like, So in a way, the whole SEO didn't have to be there because you no other people were doing it.
Mark
So basically, I'm on page one. It's just crazy if you like. Yeah. People now like how it worked back in the day. So, Yahoo, you have something similar with Yahoo. It's something similar or.
Colleen
Yeah, I think it was; I feel like I remember it having different categories. Yes, Sports and you would see all the sports websites. Yeah.
Mark
And then, if you just had a website that fit the category, I guess what you listed. People would find you, and now it's like paying for listing and then potentially not even getting to page one. It's like it's crazy. But yeah, things change. Things go so fast that I have no clue where let's say we even are in like something like 2030,
Mark
Oh yeah, it just moves so fast. I like space backgrounds. I can't wait to put a rocket on Mars as an astronaut and then go back safely.
Colleen
We just had a helicopter fly on Mars today.
Mark
Yes. Yes, I know, which is cool as well. And then, it was running Linux. There's a nerd joke in there, where like, I think you know this pun as well. I'm sorry, podcast listeners, but it's like this recurring joke of Linux on the desktop was like this year's going to be the year that Linux breaks through on the desktop.
Mark
So then there's a lot of like Linux people. And to be fair, I was one of those hardcore people as well. They were like, OK, well, maybe we're not running on the desktop everywhere. But we're flying a helicopter on Mars, like, OK, OK, I'm giving you this.
Colleen
That's a good thing. I will counter that. I just made this up. But this is my new joke. It's running Linux on Mars. It would have flown yesterday, but we had to update some drivers on that.
Mark
That's if it's running windows. Yeah.
Colleen
Yeah. Download some DLL's.
Mark
Yes. Yes, yes. But it's amazing what they do like I have something like an ARM board myself, and then it's amazing what you can do with the FPGA where it's like a small development board. And like I don't know 75 per cent of what it can do. And then you realize, like this little board that you have is like a hundred, 200 times more powerful than what they used to land on the moon before, which is like sometimes baffling like you see kids walking around like 12-year-olds or something with a phone.
Mark
You go like, the phone that you're using for, like Instagram or whatever. Yeah. That phone is like 100 times more powerful than showing a youtube video. Then the rocket they used to land. And guess what, I there's a famous picture about, I've forgotten now. The woman that wrote the code on the traction feed paper.
Colleen
Margaret Hamilton
Mark
Yes. And then you it's like you see kids looking at it like, wow paper, and then this is the funniest one I saw. And this is like a recurring meme somebody went like, oh yeah. You printed the safe icon, and it showed you like a diskette.
Colleen
Oh, God. Yeah.
Mark
And then you're like, like no, no.
Mark
I said, well, it worked. Hmm. And I remember like I had a neighbour that I used to work with, like with a PDP-11, And then he had the reels.
Colleen
Yeah.
Mark
And then you go like, well, I feel old now, but the good part my point was like this, so much cool stuff coming out still. Oh. So yeah. I am going back, going off the rails already. If a budding entrepreneur would ask you for one piece of advice, what would it be?
Colleen
Oh, I think I would say they should focus on creating an inclusive culture and hiring a diverse workforce from the start. And the reason I say that is because I see so many young startups; they lack diversity. It's just a homogenous staff. And it's easy to ignore things like this early on because you're focusing 100 per cent on the bottom line and getting a new product off the ground, finding product-market fit, etcetera.
Colleen
But early efforts in this area have a huge and lasting impact, and I think it can affect the future growth of your business. I mean, like I was saying earlier, it's been shown time and again that companies with more diverse staff and more inclusive culture perform better and innovate more. And that's regardless of the company size in the company stage. And this is an edge that you can benefit from, and it's easy to embrace. And yet, time and again, I see early-stage companies ignoring it. Then they accrue what I've seen called diversity debt, and they have to try to tackle it later.
Colleen
And it's harder to do later. And if you start with a diverse workforce and an inclusive culture, you start thinking about these things early on. I think you can get ahead in a way that many companies don't seem to realize. So if I'm talking to an early entrepreneur, that's the advice I would give them.
Mark
Yeah. And a good call. Yeah. Because just like technical depth, or security depth, it's just one of those things where it's hard to correct because you get these things where you see it now like facial recognition systems or whatever, where it ends up being biased because the data set that used to train is not representative. And then you end up with this kind of problem and then a bit of like a fun question to end this recording with. If you had a magic wand, what would you want to make happen?
Colleen
This is a really tough question, but I like a challenge. I do have a question. Am I limited in how many times I can use the wand? And can I undo things that I've done with it? So the parameters the guardrails. I need, I need the rules.
Mark
I know I should write like a FAQ for this. So so you can do one thing multiple times, or you can just do one thing and one time, and you can't undo things. So to give you an idea, other guests have said they want a covid vaccine for everyone in the world. Yeah, of course, that's it. And I can stop talking because it's like a perfect guest.
Colleen
Obviously, I want to steal that one. Yes, I would like a covid vaccine with zero side effects and instantly available and instantly effective and no booster's required. Yes. That is what I would like. So I'm going to put my parameters on that. But yeah, I mean, obviously I'd also do things like end systemic racism and get rid of the thought that people have in their heads, that a person's worth is somehow tied to the way they look or the colour of their skin because that is just frankly stupid on a galactic scale.
Colleen
So, there's one thing I think I could just, like, wave my magic wand and do. And I also would things like cure childhood illness and malaria and covid. And of course, I'd like hunger, and I'd like to buy myself a Maseratti and make sure my kids' college funds are full and pay off my mortgage. So I feel like there's like the little tangible things in my life, like a leaky faucet.
Colleen
And then there's like fixed systemic racism, and there's probably a whole spectrum in between. But yeah, those are the things that come to mind there.
Mark
Yes. There's so much stuff like, for instance, things like I, I would like to be in a position that, A: I can retire but also build schools. OK, that's kind of hard. Yeah. I just want to go to third-world countries and build schools. Yeah, totally. One of the things I thought about, like tech stuff, was like solar-powered stations, like mini-libraries/education.
Mark
Yeah, facilities, if you will. Yeah. But that they need to be rugged, of course, that kind of thing. Again if somebody wants to do that kind of startup, do call me, but yeah. It's like it's one of those things. This kind of my general point, and it's just the good thing about Covid is that remote is more normal now. There are so many people out there that are absolutely brilliant.
Mark
But because they are born in the wrong location, they might not even get to the point that somebody will listen to them, which is the sad bit. That's my hope that changes because I know there's going to be lots of people, just for instance, like bad broadband where because they have a bad Internet connection, it means that, for instance, they are not a position to, like do something simple as running webinars because their local infrastructure would not even be able to do that.
Mark
Yeah. And I'm not even talking about consistent power. I mean, like you and I can be like you know what, even if it occurs, we can be like, hey, well, let's get a generator. And like, I can run the laptop of my generator if necessary. Well, there are other countries where it'd be like, hey, if we have power two days a week, we're all very lucky, that kind of thing.
Mark
So, yeah. Removing those barriers would be great. But it's just so much to do because, let's be honest about vaccines. There are many developing and third-world countries that don't have vaccines yet because richer countries like the U.K. are now buying up all the vaccines. But the problem is, is that those other countries also need vaccines because otherwise, they become even more dependent on the richer countries.
Colleen
They don't have the distribution mechanisms the same way. And that's cool. So I will say now you feel you're making me feel bad about wanting a Maseratti with my magic wand. Can I still get my Maseratti?
Yes, yes.
Mark
I still want an Aston Martin, but I don't know if I will get one. I would say I'd like to thank you for your time.
Colleen
Of course
Mark
Of course, it was a pleasure. Hopefully for you as well.
Colleen
Absolutely yeah.
Mark
For the listeners at home, wherever you are, maybe in a submarine, in the space station. If it is, then do let me know; I would like a shout out to my daughter. Surfboards. Bali on the beach maybe in case some people are allowed to travel already. And then I will see you all next time and have a good one.
Colleen
Thank you.
Intro
Thank you for tuning in this episode. I hope you enjoyed the show, and I look forward to having you tune in again next episode. See you next time.